The Lindy Effect: Why Old Ideas Outlast New Ones

A book that has been in print for 100 years will likely remain in print for another 100. A book published last month has, statistically, roughly a month before it's forgotten.

This is the Lindy Effect: for non-perishable things — ideas, books, technologies, institutions, cultural practices — life expectancy increases with age. The older something is, the longer it's likely to last.

This sounds counterintuitive. We're trained to associate age with obsolescence, to privilege new over old, to treat recent as more relevant. The Lindy Effect says the opposite is true for a specific and interesting reason.

The Mechanism: Time as a Fragility Filter

The mechanism isn't mysterious once you understand what survival means.

Perishable things (food, human bodies, physical structures) age toward decay. A 90-year-old human has fewer expected years of life than a 30-year-old. A week-old apple has fewer edible days than a fresh one. Time increases proximity to expiration for perishable things.

Non-perishable things are different. Ideas, books, technologies, and institutions can last indefinitely — there's no biological clock. But they face a different kind of attrition: the stressors of reality. New technologies emerge that make them obsolete. New ideas challenge them. Social conditions change that undermined their relevance. Institutions face crises that collapse the ones with structural weaknesses.

What survives all of this is demonstrating something: it hasn't been killed by any of the stressors it's faced. It has encountered challenges and survived them. The survival is evidence of robustness.

When a book has been in print for 100 years, it has survived: competing books, changing academic consensus, shifts in cultural taste, the mortality of its publishers, technological changes in how books are produced and consumed, and generational turnover in readership. Whatever it contains is resilient to all of those pressures. That is information.

When a book was published last month, it has survived: getting past an editor and printer. That's it. Much less information.

Time Is the Best Expert

This is Taleb's most striking claim in this section: time is a better expert on what matters than contemporary peers.

Contemporary critics, prize committees, bestseller lists, academic citation counts, social media engagement — all of these reflect the preferences and biases of a specific moment. They capture what is currently prestigious, currently fashionable, currently alignable with institutional incentives.

They do not reliably track what is genuinely valuable across time and conditions. Countless books won prizes and were celebrated as transformative and are now forgotten. Countless ideas were dismissed as eccentric or marginal and are now foundational.

The Lindy test cuts through this noise. Not: is this celebrated now? But: has this survived for how long, under what conditions?

"Books of the type written by the current hotshot Op-Ed writer at the New York Times may get some hype at publication time, but their five-year survival rate is generally less than that of pancreatic cancer."

Worked Examples

Books: Shakespeare has been in continuous production for 400 years. The mechanism is Lindy: each decade he survived is evidence that something in the work resonates across contexts that its author couldn't anticipate. A novelist celebrated for capturing "the zeitgeist" of their precise moment is betting against Lindy — zeitgeists don't last.

Cooking: Ancient recipes from Persia, China, Italy, and India are still eaten because they work. They have been tested against real human palates across thousands of years and survived. The trendy restaurant dish invented last year hasn't been tested by anything except the current fashion.

Medical practices: The advice "wash your hands before surgery" is centuries old and survived on results. Bloodletting was contemporary expert consensus and failed the Lindy test — it produced observable harm and eventually died. The survival of hand-washing through the development of germ theory, through mechanized hospitals, through a century of modern medicine, is real information.

Technology: The hammer, the wheel, the book (codex form) — these are ancient technologies still in daily use. The mobile app launched last year hasn't been tested by time. New doesn't mean better; old doesn't mean obsolete.

Lindy-Compatible Heuristics

Taleb derives several practical heuristics from the Lindy Effect:

Trust old books over new ones in proportion to the length of their survival. This doesn't mean only reading old books. It means calibrating your confidence in the ideas proportionally. A 2,000-year-old ethical framework has survived philosophical challenges and real-world testing that the framework proposed last year hasn't.

Be more skeptical of new dietary advice than of traditional eating patterns. Traditional diets have been tested by survival across generations of real human health outcomes. New dietary theories have been tested by studies with small samples and short timeframes.

Prefer old technology for critical functions when new technology is less proven. The proven technology has demonstrated reliability across more failure modes. The new technology may be better on the dimensions it was designed to optimize; it hasn't been tested on the full range of conditions it will eventually face.

Give weight to folk wisdom and grandmother's advice that has survived generations. It wasn't transmitted because it was wrong. The selection process is evidence.

The Limit of Lindy

The Lindy Effect isn't universal. Some things are old because they were monopolized by institutions with the power to sustain them regardless of their value. Some old institutions survive not through demonstrated value but through entrenched power.

The test isn't just age; it's survival under competitive pressure. Ideas that survived despite active attempts to displace them are more Lindy-robust than ideas that survived because they were protected.

For the full framework, read Skin in the Game Explained.